37.2 Overview of Technical Terms
This section provides a brief overview of some of the commonly used technical terms that are relevant in the field of indoor localization [1].
Absolute and Relative Location. A location determined within the context of a global or large area reference grid obtained from GNSS satellites, markers, or landmarks is referred to as an absolute location. In contrast, relative positions depend on a local frame of reference, for example, coordinates within a small coverage area that represent displacement with respect to a local fixed reference (e.g. a fixed Wi‐Fi access point with known global coordinates).
Anchor and Mobile Nodes. From a networking perspective, nodes in indoor environments that are part of the network and have a stationary (fixed) location are referred to as anchor nodes. In the literature, such nodes may also be referred to as beacons, fixpoints, access points (APs), base stations, or reference nodes. Typically, the coordinates of such anchor nodes are assumed to be known. In contrast, nodes that are part of the network and can move in the indoor environment are referred to as mobile nodes. Such nodes could represent people, robots, or other locomotion‐capable devices (e.g. drones). In general, it is the job of the localization system to determine the (local or global) coordinates of such mobile nodes.
Centralized and Distributed Localization. In a centralized localization architecture, location estimation is carried out at a central server where all anchor and mobile node locations are stored and available to an administrator. The benefits of centralized architectures are simplicity, uniform service to all users, and lower expansion costs as most of the intelligence in the system is concentrated at one location, allowing the mobile and anchor nodes to be lower cost and contain fewer components. In a distributed system, location estimation is carried out on each mobile and anchor node based on local observations. The advantages of a distributed architecture are good system scalability and better guarantees of the user’s privacy (as sensitive location information is not centrally stored, making it less susceptible to being compromised).
Line of Sight (LOS). When a signal can travel via a direct straight path from an emitter to a receiver, it is referred to as LOS transmission. Several localization techniques rely on LOS, for example, time of arrival (ToA)‐based distance measurements with radio frequency (RF) signals. But due to occlusions from walls, furniture, and people, most indoor environments typically induce non‐LOS (NLOS) propagation, which may cause inconsistent time delays at a radio receiver. These delays pose a challenge that can only be tackled by few localization techniques.
Multipath Environment. An environment in which a transmitted signal propagates along multiple paths (echoes), each of which arrives with different path delays at the receiver, is referred to as a multipath environment. Multipath propagation of signals is particularly problematic for time‐based localization methods (Section 37.5.1.2) because signal paths from different directions degrade the ability to determine the travel time of the direct path. One way to distinguish the direct path from a non‐LOS path is to move the receiver or transmitter. Non‐LOS paths change erratically while in motion, allowing for separation and averaging, while the direct path is directly related to the motion of the object. Thus, averaging over time with a motion‐tracking model is one effective way to mitigate multipath. Another way to overcome multipath is to switch to different frequency channels. Alternatively, radio signals with a large absolute frequency bandwidth such as Ultra‐Wideband (UWB) have been shown to be advantageous for mitigation of multipath fading [2].
Received Signal Strength Indicator (RSSI). Signal attenuation can be used for distance estimation during localization, based on RSSI values. RSSI are observed RSS (received signal strength) values averaged over a specific sampling period and usually specified as received power PR in decibels. Based on the attenuation model
the received signal power or signal strength PR can help with the estimation of the distance d of a mobile user or object from the transmitter. In this model, PT is the transmitted power at the transmitter, GT and GR are the antenna gains of the transmitter and receiver, and p is the path loss exponent. The path loss factor p characterizes the rate of attenuation with an increase in distance d. The free space model does not take into account that antennas are usually set up above the ground. In fact, the ground acts as a reflector, and thus the received power differs from that of free space. A mathematical formulation of such a path loss model, also known as open field model, can be found in [3]. Typically, in free space p = 2, whereas for environments with NLOS multipath, p > 2. For indoor environments, the path loss exponent typically takes higher values between 4 and 6. Theoretically, distances estimated from RSSI values to multiple anchor nodes can be used to determine the receiver position by multilateration techniques (see Section 37.5.1 for more details). However, interference, multipath propagation, and presence of obstacles and people results in a complex spatial distribution of RSSI values, which can make the estimation of distances using RSSI alone quite inaccurate. Therefore, fingerprinting has become more popular than propagation modeling (see Section 37.5.2 for more details).
37.3 Performance Metrics
Indoor localization solutions need to meet several goals if they are to be considered viable candidates for use in indoor environments. Here we review some of the more relevant performance metrics [4] that must be satisfied by any candidate indoor localization solution:
Accuracy: The location error of a positioning system is one of the most important metrics used to determine the effectiveness of a localization system. In its simplest form, localization accuracy can be reported as an error distance between the estimated location and the actual location of the user or object being tracked. For navigation systems, this may take the form of a running average of errors over a time period of interest, or the error could be calculated using geometric principles, to estimate the deviation of the predicted trajectory from the actual trajectory. Usually, the higher the accuracy, the better the system, but there is often a trade‐off between accuracy and other characteristics. Therefore, a compromise between adequate accuracy and other characteristics described below is essential.
Timeliness: The timeliness or responsiveness of a solution determines how quickly the location estimate of a target is obtained. For simple indoor localization queries, a fast response to the query is important in most cases, but not crucial. However, for navigation systems, timeliness is a critical measure of effectiveness: if location estimates are not updated quickly in sync with the motion profile of the subject being tracked, the system will be ineffective for the purpose of navigation (regardless of the eventual accuracy of the estimates). Usually, the term location lag is used to refer to the delay between a mobile subject moving to a new location and the new location of that subject being reported by the system.
Coverage: Any indoor localization solution must work and be usable over the entire indoor environment of interest. Coverage defines the area over which a localization solution can provide estimates of sufficient accuracy, and possibly timeliness, to be considered useful. The physical environment (e.g. obstacles, walls, doors) plays a crucial role in limiting the availability of signals that are used by a given localization technique, consequently impacting the coverage achievable by the technique for that environment. Intuitively, it is possible to extend coverage by altering the physical environment or supplementing it with additional hardware, for example, wireless signal repeaters. Coverage can also be improved by enhancing the hardware carried by the user or object being tracked, for example, using mobile devices with more powerful and capable wireless radio